GRAPES青藏高原混合同化方法初步试验
投稿时间: 2016-01-19  最后修改时间: 2016-03-22  点此下载全文
引用本文:
摘要点击次数: 153
全文下载次数: 
作者单位E-mail
夏宇 成都信息工程学院 245171196@qq.com 
基金项目:国家自然科学基金项目(面上项目,重点项目,重大项目) ;国家重点基础研究发展计划(973计划)
中文摘要:摘要:针对青藏高原地区气象观测站点稀少和模式同化分析质量较低问题,将GRAPES区域集合变分(En-3DVAR)混合同化系统应用于青藏高原地区,进行高原单点理想试验和真实观测资料同化分析预报试验,分析评估青藏高原混合同化分析增量及预报误差的水平垂直结构特征及其合理性,并与中国东部平原地区进行比对,以期通过混合同化方法改进青藏高原数值预报同化分析场质量和模式预报效果。单点理想试验表明:混合同化系统中风场和气压场在平原地区和青藏高原地区都满足地转平衡约束关系,En-3DVAR系统与3DVAR系统相比,出现了一些小尺度分析增量,特别是在高原地区表现更明显。真实资料混合同化试验结果表明:集合预报估计的背景误差协方差具有流依赖的特征,特别是在高原地区和低值系统附近;青藏高原地区分析场绝对误差总体低于3DVAR系统,青藏高原分析场平均绝对误差在中层和高层低于平原地区,说明青藏高原地区改进效果略优于平原地区。需要关注的是,青藏高原地区En-3DVAR混合同化分析增量总体大于3DVAR,特别是近地面层u风分量分析增量明显偏大,这可能与青藏高原复杂地形有关,需要进一步改进集合预报,进而改进集合估计的背景误差协方差。
中文关键词:GRAPES,青藏高原,混合同化方法,初步试验
 
The tentative experiment of GRAPES En-3DVAR hybrid data assimilation method in Tibet platrau
Abstract:Abstract:In view of the sparse observation stations and low quality of the model analysis in the Tibet Plateau,application the Ensemble-3DVAR hybrid data assimilation system (En-3DVAR) to the Tibet plateau, a single pressure observation data assimilation experiment and a real observation data of assimilation analysis experiment and prediction experiment is conducted, analyse and evaluate the horizontal and vertical structure and rationality of the Tibetan Plateau En-3DVAR hybrid assimilation analysis increment , and compared them with the plain area of Eastern china, in order to improve the quality of the data assimilation analysis field of the Tibetan Plateau and increase the effect of the model by using the En-3DVAR hybrid assimilation method. The single pressure observation data assimilation experiment showed that both in the Tibet Plateau and plain are satisfied the wind and pressure geostrophic equilibrium constraint relation, compared with 3DVAR,En-DVAR has revealed some small scale analysis increment, especially in the Tibet plateau,the performance is more obvious. The real data En-3DVAR hybrid assimilation experiment showed that the background error covariance of ensemble forecasting has the characteristics of flow dependence,especially in Tibet plateau and low value system;The absolute error of the analysis field in Tibet Plateau is generally lower than 3DVAR system,the mean absolute error of the analysis field in the Tibet Plateau is lower than plain in the middle and high level,the result shows that the improvement of the Tibet plateau is better than plain area; It need our attention that in the Tibet Plateau the analysis increment of the En-3DVAR hybrid assimilation is generally greater than 3DVAR,especially on the surface layer, the u-wind component analysis increment is obviously too larger than 3DVAR,this may be related to the complex terrain of the Tibet Plateau,witch need improve the ensemble forecast system to optimize it background error covariance.
keywords:GRAPES, Tibet plateau, hybrid data assimilation, tentative experiment
查看全文  查看/发表评论  下载PDF阅读器